Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

about CropResizeRotate #41

Open
xiaomingdaren123 opened this issue Sep 12, 2018 · 4 comments
Open

about CropResizeRotate #41

xiaomingdaren123 opened this issue Sep 12, 2018 · 4 comments

Comments

@xiaomingdaren123
Copy link

hi,MarekKowalski ,In the CropResizeRotate,I don't understand the role of destShape.Why is this step to calculate the transformation from initshape to destshape? In the network,the initLandmarks(S0) is the destshape?thanks

@MarekKowalski
Copy link
Owner

Hi,
The destShape is the shape you want your input shape to be warped to (using a similarity transform).

In the network destShape is initLandmarks, which is a frontal face shape located in the middle of a 112x112 image. All input images are warped so that their initial landmarks are aligned with those of initLandmarks.

Does that answer your question?

Thanks,

Marek

@xiaomingdaren123
Copy link
Author

Thank you very much for your answer.

@gmlyytt-YANG
Copy link

gmlyytt-YANG commented Dec 23, 2018

Hello @MarekKowalski , Thank you very much sharing the code of your work. Could you explain a little about the role of frameFraction in ImageServer? Is that the true imgSize of the real application of your dan model?

Young
Thank you very much

@MarekKowalski
Copy link
Owner

MarekKowalski commented Dec 31, 2018

Hi Young,

frameFraction refers to the fraction of the input image that will be the "frame", which is the area outside of the landamarks of initLandmarks. If you set this value to be high, the face will take a smaller part of the image and vice versa.

I'm not sure what you mean by true imgSize :(

Thanks,

Marek

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants